So far, we have considered statistical methods that consider only one (univariate) or two variables (bivariate) at a time. The problem with this is that the social world is multivariate in nature. That is, the causes of the social phenomena that social scientists are interested in studying are many. Take for example a researcher interested in sentencing disparities based on race. The relationship between race and sentence length could be examined (assuming that accurate data can be obtained), but this picture would be far from complete unless the seriousness of the offense was taken into account. If the researcher were to include the seriousness of the offense and the race of the defendant, then a more complete picture will emerge. The relative effects of both variables can be examined.
Our last chapter, then, considers statistical techniques that consider sets of variables and how different sets of variables relate to each other as sets. Because there are more than two variables involved, these methods are usually classified as multivariate statistics. We will stay away from the complex mathematics necessary to obtain the results of these methods. Our concern will be with the logic of particular methods, the assumptions, and the interpretation of results. The hope is that when you finish this chapter, you will be able to read and interpret the social scientific literature that presents the results of these methods.
Students: If you are using this book for an undergraduate course, pay special attention to your professor’s instructions regarding this material. It may be considered too advanced, and you may not have time to cover it in a single term. You will most likely be assigned parts of this material rather than all of it. If you are a graduate student, then most likely you will have to master all of the concepts in this section.
Last Modified: 07/02/2018